Sie sind auf Seite 1von 8

Renewable Energy 76 (2015) 679e686

Contents lists available at ScienceDirect

Renewable Energy
journal homepage: www.elsevier.com/locate/renene

A dynamic thermal performance model for flat-plate solar collectors


based on the thermal inertia correction of the steady-state test
method
Jie Deng a, b, Yupeng Xu a, Xudong Yang a, *
a
Department of Building Science, Tsinghua University, Beijing 100084, PR China
b
Institute of Electrical Engineering, Chinese Academy of Sciences, Beijing 100190, PR China

a r t i c l e i n f o a b s t r a c t

Article history: In determining the dynamic thermal performance of a flat-plate solar collector, when the instantaneous
Received 28 July 2014 solar irradiance changes sharply at one moment, most of the existing models cannot accurately predict
Accepted 1 December 2014 the momentary thermal characteristics of outlet temperature and useful heat gain. In the present study,
Available online 12 December 2014
an analytical model in the form of series expansion is put forward to depict the momentary thermal
characteristics of flat-plate solar collectors. The analytical model reveals that, instantaneous useful heat
Keywords:
gain of a solar collector at one moment consists of the steady-state useful heat gain and corresponding
Flat-plate solar collector
thermal inertia correction. The model is then validated by the experimental data. It indicates that the
Dynamic thermal performance model
Steady-state test
analytical model can properly predict the dynamic thermal performance of the solar air collector. Be-
Thermal inertia sides, the model pertains to other types of solar thermal collectors, if they can be tested by the steady-
state test method.
© 2014 Elsevier Ltd. All rights reserved.

1. Introduction purpose of product's quality control. However, the NDM requires


familiarity with digital signal analysis and filtering [3]. Thus the
Flat-plate solar collectors are popularly used for providing calculation by NDM is very complex. When it comes to QDT, the
domestic hot water or space heating. In the practical applications, model is just a simplified quasi-dynamic model of a flat-plate solar
solar collectors usually work under dynamic weather conditions. So collector with a correction term of thermal capacitance, which is
the thermal performances of the solar collectors are dynamic and inferior to the quasi-dynamic test method (QTM) by Perers [7e10]
the steady-state model derived from steady-state test (SST) method from the view point of theoretical completeness. Furthermore,
[1] does not pertain to the dynamic conditions. Hence, dynamic Kong et al. [18] verified that the transfer function method (TFM)
model is needed to predict the thermal performances of flat-plate [19,20] is more accurate than the QTM by Perers. It seems the TFM
solar collectors under realistic weather conditions. Researchers has a good accuracy than the QDT and QTM methods. However,
have developed dynamic test methods for flat-plate solar collec- dynamic thermal performance predictions done by Kong et al.
tors. Amer et al. [2] and Nayak and Amer [3] gave a detailed review [18,21] with flat-plate solar collectors suggest that, the TFM cannot
of transient methods, such as one-node method [4e10], multi-node accurately predict the momentary thermal characteristics of outlet
method [11,12], response function method [13e16], etc. The dis- temperature and useful heat gain when the instantaneous solar
advantages of most of the one-node method are described by Amer irradiance changes sharply at one moment.
[2]. Nayak and Amer [3] carried out sensitivity study to examine the In order to accurately predict the dynamic thermal performance
effect of uncertainties in measurements on the values of the esti- of a flat-plate solar collector when the instantaneous solar irradi-
mated parameters using different methods. They found that the ance changes sharply, the present study puts forward an analytical
new dynamic method (NDM) [2,16] seems to be quite reliable and model in the form of series expansion, starting from the simplified
the quick dynamic test (QDT) [4,17] method is the simplest for the model of QDT [4,17] of a flat-plate solar collector. Based on the
analytical model deduced in the condition of a small time interval,
the physical explanation of the model is presented and the solving
* Corresponding author. Tel.: þ86 10 6278 8845; fax: þ86 10 6277 3461. method is elaborated. Moreover, experiments are conducted to
E-mail address: xyang@tsinghua.edu.cn (X. Yang). validate the accuracy of the analytical model. And the dynamic

http://dx.doi.org/10.1016/j.renene.2014.12.005
0960-1481/© 2014 Elsevier Ltd. All rights reserved.
680 J. Deng et al. / Renewable Energy 76 (2015) 679e686

Nomenclature SST steady-state test


T response time constant, s
Aa transparent frontal area or the aperture area of a TFM transfer function method
collector, m2 TICM thermal inertial correction model
Ag gross collector area, m2 Ta ambient temperature,  C
b0 constant used in incident angle modifier equation, Tf characteristic temperature of the working fluid,  C
dimensionless Tfi collector inlet temperature,  C
Cf specific heat of the working fluid, J/(kg  C) Tfo collector outlet temperature,  C
FR solar collector heat removal factor, dimensionless Tfo,initial initial collector outlet temperature in the test of time
f (n) useful heat gain of the collector in the steady-state test constant,  C
method, W UL overall heat loss coefficient of a solar collector, W/
Gb beam irradiance of inclined surface, W/m2 (m2  C)
Gd diffuse solar irradiance of inclined surface, W/m2
Gg global solar irradiance of inclined surface, W/m2 Greek symbols
j a forced integer number in Equations (16) and (17), a absorptance, dimensionless
dimensionless b collector slope angle, 
Kqb(q) collector incidence angle modifier for beam irradiance, hg collector thermal efficiency based on the gross area, %
dimensionless q incidence angle on the tilted surface of a collector, 
Kqd(q) collector incidence angle modifier for diffuse r density, kg/m3;
irradiance, dimensionless t time, s; transmittance of glass cover, dimensionless
m mass, kg (ta)en effective transmittance-absorptance product at normal
(mC)e effective thermal capacitance of solar collector, J/ C incidence, dimensionless
·
mf mass flow rate of flowing air, kg/s Dt small time interval, s
N number of time step, dimensionless
NDM new dynamic method Subscripts
QDT quick dynamic test a ambient
QTM quasi-dynamic test method f working fluid
Qu instantaneous useful heat gain of the collector i inlet air
dynamic test, W o outlet air
S absorbed solar radiation per unit area, W/m2

thermal performance predictions by different models are compared with inclined angles of 0e90 . The calculation method of the
with the experimental data. In addition, the applicable condition of equivalent angle of incidence can be found in Ref. [17].
the analytical model is discussed.
   
S ¼ ðtaÞen Kqb q Gb þ ðtaÞen Kqd q Gd (2)

 
2. Analytical model 1
Kqb ðqÞ ¼ 1  b0 1 (3)
cos q
2.1. Transformation of simplified quasi-dynamic model
The model of QDT by Equation (1) represents a simplified first-
The simplified model of QDT [4,17] for a flat-plate solar collector order differential system, which has a mathematical form of
is given in Equation (1). It is a first-order differential model based Equation (4), as verified by Ref. [22].
on Hottel-Whillier-Bliss equation [17] with a correction term of
effective collector thermal capacitance [22]. dy
T þ y ¼ Ax (4)
dt
dTf h  i ·  
ðmCÞe ¼ FR Aa S  UL Tfi  Ta  mf Cf Tfo  Tfi (1) where y is the system output, x is the system input, T is the time
dt
constant of the object, A is the system amplification coefficient.
where the term on the left of the equation represents the effective Similar to the form of Equation (4), taking the collector outlet
internal energy variation. The first term on the right indicates the temperature Tfo as the system output of the solar thermal collector
effective absorbed radiation subtracting the total heat loss in terms and substituting the characteristic fluid temperature Tf with
of the heat removal factor FR. The second term on the right denotes (Tfi þ Tfo)/2, Equation (1) can be rearranged to Equation (5).
the thermal enthalpy increment between the collector out and
·
inlet. (mC)e is the effective heat capacity of the solar collector, J/ C. dTfo 2mf Cf   dTfi 2FR Aa h  i
The absorbed radiation per unit area (S) consists of beam radiation þ Tfo  Tfi ¼  þ $ S  UL Tfi  Ta
dt ðmCÞe dt ðmCÞe
and diffuse radiation [7e10], as given in Equation (2). The incidence
(5)
angle modifier Kqb(q) for beam radiation is given by Equation (3).
With regard to the incidence angle modifier Kqd(q) for diffuse ra- In the test process of response time constant T of a solar col-
diation, an equivalent angle of incidence can be considered ac- lector, the temperatures Tfi,Ta, outdoor wind speed, and the mass
·
cording to Duffie and Beckman [17]. And Kqd(q) is the cosine value of flow rate mf of the working fluid are kept to be constant. The term
dTfi
the equivalent angle, ranging from 55 to 60 for solar collectors dt
is zero in this case. Then Equation (5) is reduced to Equation (6)
J. Deng et al. / Renewable Energy 76 (2015) 679e686 681

and the expression of response time constant T of the collector recurrence relation, if n > 3. The recursive process is shown in
outlet temperature is given by Equation (7). Equation (14) and the final expression of Qu(n þ 1) is given by
Equation (15), which is in the form of a series expansion.
dTfo   F A h
a
 i
 
T þ Tfo  Tfi ¼ · R $ S  UL Tfi  Ta (6) Dt
dt m C Qu ðn þ 1Þ ¼ f ðnÞ þ ½Qu ðnÞ  f ðnÞ$exp  (12)
f f
T
where T is the response time constant of a collector.
where
ðmCÞe h  i
T¼ · (7) f ðnÞ ¼ FR Aa $ SðnÞ  UL Tfi ðnÞ  Ta ðnÞ (13)
2mf Cf
    
Dt Dt
Qu ðn þ 1Þ ¼ f ðnÞ$ 1  exp  þ Qu ðnÞ$exp 
T T
2.2. Analytical model in the form of series expansion      
Dt Dt
¼ f ðnÞ$ 1  exp  þ f ðn  1Þ$ 1  exp 
Equation (5) is a transformation of the quasi-dynamic model of T T
   
Equation (1). And Equation (6) is a reduced energy balance equa- Dt 2Dt
dT $exp  þ Qu ðn  1Þ$exp 
tion of Equation (5), which neglects the differential term dtfi . In fact, T T
     
when the time interval is small, the collector inlet temperature Tfi Dt Dt
dT
can be assumed to be constant. Then dtfi equals zero and Equation ¼ f ðnÞ$ 1  exp  þ f ðn  1Þ$ 1  exp 
T T
(6) can be used to model the thermal performance of a flat-plate     
Dt Dt
solar collector. Hence, we deduce a new expression of instanta- $exp  þ f ðn  2Þ$ 1  exp 
T T
neous useful heat gain Qu in light of Equation (6) based on a small    
2Dt 3Dt
time interval as follows. $exp  þ Qu ðn  2Þ$exp 
Firstly, take integer n ¼ 1,2,3,… as an arbitrary time step and T T
Tfi(n),Tfo(n), Ta(n), S(n) represent according parameters at the time (14)
·
n$Dt. Here the time interval Dt for Tfi, Ta, S, mf is uniform. In each
·   
time interval Dt, the values of Tfi, Ta, S, mf and consequently FR are X
n
Dt
constant values for integrating Equation (10). The single time- Qu ðn þ 1Þ ¼ f ðn þ 1  iÞ$ 1  exp 
T
depending variable is Tfo. When the time interval Dt between i¼1
    (15)
the moment n$Dt and (n þ 1)$Dt is small, the collector outlet ði  1ÞDt nDt
$exp  þ Qu ð1Þ$exp 
temperature Tfo(n þ 1) at time (n þ 1)$Dt can be solved by def- T T
inite integral. The process of solving is given by Equations
(8)e(11).
( )
dTfo 1   FR Aa h  i 2.3. Model explanation
¼ Tfo  Tfi  · $ S  UL Tfi  Ta (8)
dt T mf Cf
From Equation (13), the formula of f(n) represents the steady-
state useful heat gain of a collector, which is generally known by
dTfo dt the steady-state test (SST) method. Uniformly, in the analytical
  h  i ¼  (9)
Tfo  Tfi  FR Aa
· $ S  UL Tfi  Ta T model given by Equation (15), the terms f(n þ 1  i) represent the
mf Cf
steady-state useful heat gain. And the exponential functions
Integrating on both side of Equation (9) from the time n$Dt to exp(iDt/T) (i ¼ 1, 2, …, n) indicate the thermal inertia effect of the
(n þ 1)$Dt, we have. collector, which can be illustrated by the test of the response time
constant and will be further discussed in section 3.1. It suggests that
ðnþ1Þ
Z ðnþ1Þ$Dt
Z the analytical model is essentially a thermal inertial correction
dTfo dt
  h  i ¼  (10) model (TICM) of the SST method.
Tfo  Tfi FR Aa
 · $ S  UL Tfi  Ta T Moreover, when nDt  3T at the time step n, the thermal inertia
n mf Cf n$Dt
effect term is very small and can be neglected, since exp(3T/
h i F A h  i T) ¼ 0.0498. Therefore, the useful heat gain Qu ðn þ 1Þ at time
a
Tfo ðn þ 1Þ  Tfi ðnÞ  · R $ SðnÞ  UL Tfi ðnÞ  Ta ðnÞ t ¼ ðn þ 1Þ$Dt is mainly affected by the time interval of ½t  3T; t.
mf Cf And thus the TICM can be expressed as Equation (16), based on the
( )
h i F A h  i SST and thermal inertia effect.
R a
¼ Tfo ðnÞ  Tfi ðnÞ  · $ SðnÞ  UL Tfi ðnÞ  Ta ðnÞ
mf Cf ðjþ1Þ
X   
  Dt
Dt Qu ðn þ 1Þ ¼ f ðn þ 1  iÞ$ 1  exp 
$exp  i¼1
T
T    
ði  1ÞDt ðj þ 1ÞDt
(11) $exp  þ Qu ðn  jÞ$exp 
T T
Instantaneous useful heat gain Qu(n þ 1) at time (n þ 1)$Dt can (16)
·
be expressed as Equation (12) with Tfi ðn þ 1Þ¼ Tfi ðnÞ, by multi-
·
plying mf Cf on both sides of Equation (11) and rearranging. Equa- where the rounded off integer j ¼ ½3T=Dt.
tion (12) is a recurrence formula which depicts the relation of the The term Qu ðn  jÞ$exp½ðj þ 1ÞDt=T in Equation (16) can be
instantaneous useful heat gain Qu(n þ 1) and Qu(n). Therefore, neglected due to the mathematical relation given by Equation (17),
Qu(n þ 1) is related to Qu(n), Qu(n  1), Qu(n  2),…,Qu(1) by considering that Qu ðn  jÞ holds the order of magnitude of
682 J. Deng et al. / Renewable Energy 76 (2015) 679e686

Qu ðn þ 1Þ. So the TICM is truncated to Equation (18), which is used


in the following analysis.
   
ðj þ 1ÞDt ð3T=Dt þ 1ÞDt
exp   exp   expð3Þ
T T
¼ 0:0498
(17)

ðjþ1Þ
X   
Dt
Qu ðn þ 1Þ ¼ f ðn þ 1  iÞ$ 1  exp 
T
i¼1
  (18) Fig. 1. Schematic of the solar air collector test rig. 1 e Centrifugal fan; 2 e Vent valve; 3
ði  1ÞDt e Soft steel duct; 4 e Steel air conduit; 5 e Rotating rack; 6 e Intelligent air flowmeter;
$exp  ; j  ½3T=Dt
T 7 e Thermocouples; 8 e Pyranometer; 9 e Flat-plate solar air collector; 10 e Collector
inlet; 11 e Collector outlet; 12 e Pressure drop transducer; 13 e Electrical Heater; 14 e
Thermometer screen.

2.4. Solving method


outlet. A thermometer screen was used to obtain the ambient
The time constant T can be obtained by the response of collector temperature. All the instruments were calibrated before experi-
outlet temperature when the collector is abruptly shielded from the ments. The exit mass flow rate m_ e of the flowing air through the
sunshine with an opaque cover. T is the time required for the interior of the collector was controlled to be near constant, viz.,
quantity ðTfo;t  Tfi Þ=ðTfo;initial  Tfi Þ to change from 1.0 to 0.368 [1]. 0.0405 kg/s with a deviation of ±2%. The time interval of data
As an alternative, T is the time required for the fluid leaving a solar acquisition was 10 s. Besides, the wind speed ranged from 0 to 2 m/s
collector to attain 63.2% of its steady-state change following a step during the test, measured by a portable anemometer at random
change in irradiance, according to ASHRAE 93-2003 [1]. The pa- times.
rameters UL , FR can be obtained by least squares linear regression
using SST method. The parameters m_ f , S, Tfi , Ta are measurable 3.1. Steady-state test (SST)
quantities. And the effective thermal capacitance ðmCÞe of the
collector is calculated by Equation (19), which is equivalent to A minimum of 16 data points at four different inlet tempera-
Equation (7). Hence, the coefficients in the TICM given by Equation tures were obtained using SST method. Fig. 3 shows the thermal
(18) can be obtained according to the SST method and a single test efficiency curve hg vs. ðTfi  Ta Þ=Gg of the SAC. The error bar of each
of the collector time constant T. measurement data point is given by error analysis method [22].
· Fig. 4 gives the response curve of the collector temperature dif-
ðmCÞe ¼ 2mf Cf $T (19) ference ðTfo  Tfi Þ after irradiance interruption by means of col-
lector shadowing. The time constant T of the SAC is 9 min (540 s).
When the TICM is used to predict the dynamic thermal perfor-
The exponential relation of ðTfo;t  Tfi Þ versus time t is demon-
mance of a flat-plate solar collector, the formula in the form of
strated by Fig. 5. It denotes that the thermal inertia effect of the
series expansion given by Equation (18) can be calculated by pro-
SAC is close to an exponential relation with time, as mentioned in
grams of high-level languages on a computer, such as Matlab,
section 2.3. From Fig. 5, the relation of the quantities
FORTRAN, C language, etc. Moreover, there is a more convenient
lnðTfo;t  Tfi Þ=ðTfo;initial  Tfi Þ versus time t is not strictly a linear
way to calculate Equation (18). One can use the recurrence formula
relationship. Because the system of a collector is an n-order system
of Equation (12) to precisely calculate the instantaneous useful heat
and a first-order differential model is just a simplification, as
gain Qu ðiÞ for each time step by iterations using Excel sheet.
argued by Hou et al. [22].
3. Experimental procedure
3.2. Dynamic test
Experiment was conducted with a solar air collector (SAC) in
Beijing, China. Both steady-state and dynamic tests of the SAC were Dynamic test was conducted on the day of July 6th, 2013. The
performed. And a single test of the response time constant T was SAC was tilted due south during the test and the fan was started at
done. Schematic diagram of the SAC test rig is given in Fig. 1. And 8:31 am and stopped at 16:38 pm. Fig. 6 gives the meteorological
Fig. 2 is the picture of the test system. The SAC has one transparent
cover and its tilted angle is 45 . The gross collector area Ag is
1.985 m2 with a contour size of 1.995 m by 0.995 m
(length  width). Aperture area Aa of the collector is 1.896 m2. The
total thickness of the SAC is 135 mm. In the test process, global solar
irradiance, diffuse irradiance, inlet and outlet volume rate of
airflow, inlet temperature, outlet temperature, ambient tempera-
ture were all measured. Two pyranometers were respectively used
to measure the global solar irradiance and diffuse irradiance, with
the measuring accuracy of ±1%. The volume rates of airflow at the
collector inlet and outlet were measured by intelligent flow meters
with an accuracy of ±1.0%. Temperatures of the observation points
were measured by thermocouples of 4RTD-type, with the accuracy
of ±(0.3 þ 0.005  jTj) ( C). There were five measuring points
distributed at each cross section of the solar air collector inlet and Fig. 2. Test rig of a solar air collector.
J. Deng et al. / Renewable Energy 76 (2015) 679e686 683

Fig. 5. Exponential curve fitting of ðTfo;t  Tfi Þ=ðTfo;initial  Tfi Þ versus time.
Fig. 3. Thermal efficiency curve of the solar air collector with error bars.

collector system [23]. Moreover, the time required for the quantity
conditions of global solar irradiance Gg , diffuse solar irradiance Gd ,
Qu ðnÞ=Qu ð1Þ to change from 1.0 to 0.368 is 530 s, which is close to
ambient temperature Ta , the collector inlet temperature Tfi and
the time constant of 540 s. It suggests the TICM can catch the
outlet temperature Tfo . As in Fig. 6(b), Tfo immediately increased to
thermal inertia effect of the SAC.
a peak value of 56.2  C and then fell down gradually to 42.0  C at
8:49 am. It was because the SAC was in open-air basking before
8:31 am and thus the collector itself had some energy storage ca- 4.2. Dynamic thermal performance prediction
pacity. When the fan was started at 8:31 am, the heat stored by the
SAC gradually released until the steady-state point occurred at Based on the solving method mentioned in section 2.4, the pa-
8:49 am. Then Tfo gradually increased after 8:49 am, since twice the rameters in the TICM described by Equation (18) is shown is Table 1.
response time constants (18 min) had elapsed and the collector The incidence angle modifier coefficient b0 of the SAC is 0.1142 by
thermal inertia dominated the momentary heat storage quantity of extra test, so Kqb ðqÞ can be calculated by solar incidence angle q at
the collector. different time. Instantaneous solar incidence angles can be calcu-
lated by the formula in the reference [17], with the latitude of 39.8
4. Model verification and discussion in Beijing. And then the analytical model by Equation (18) is used to
predict the dynamic changes of instantaneous useful heat gains on
4.1. Response time constant of the collector

The response curve of instantaneous Qu during the test process


of the collector time constant also can be used to depict the thermal
inertia effect of the collector. With the initial quasi-steady state
useful heat gain Qu ð1Þ, setting subsequent solar irradiance as zero,
f ðnÞ in Equation (12) equals zero. Then the relation between Qu ðn þ
1Þ and Qu ðnÞ reduces to be exponential. Fig. 7 gives the comparison
of instantaneous Qu between the response curve by test and the
prediction by TICM during the collector time constant test period.
Good agreement is obtained between the prediction and the test
before 600 s with a relative error of ±5%. For the time after 600 s,
the relative errors between the predicted and measured Qu become
larger. This is due to the first-order simplification of the n-order

Fig. 4. Response curve of the collector temperature difference ðTfo  Tfi Þ. Fig. 6. Dynamic test of the SAC on July 6th, 2013 in Beijing.
684 J. Deng et al. / Renewable Energy 76 (2015) 679e686

Fig. 7. Comparison of instantaneous Qu between the response curve by test and the
prediction by TICM.

Fig. 8. Comparison of instantaneous Qu between the experimental data and the pre-
diction by SST method.
the day of July 6th, 2013. In order to validate the accuracy of the
TICM, the comparisons of experimental data with the predictions
by the SST model, the second-order differential TFM [21] and the Fig. 8 are considerably underestimated. Specially, an abrupt varia-
TICM are conducted. In the prediction by TICM, the iterative tion of useful heat gain to zero is predicted by the SST model after
calculation started at an initial condition of steady-state. Namely, irradiance interruption by means of collector shadowing. But in
the initial Qu ð1Þ at the time 8:31 is equal to f ð1Þ, then the thermal reality, the heat gains gradually fall off as shown in Fig. 7.
inertial effect is considered in the following values of Qu ðiÞ (i ¼ 2, 3, Regarding the second-order differential TFM, the test conditions
4, …, n). are not so strict compared to the SST model. The test condition of
Figs. 8e10 give the comparisons of predictions by the three global irradiance is broadened to 300 W/m2 [21]. Although the TFM
models with the experimental data, respectively. Since the SAC was considers the collector capacitance of the solar part and the fluid
in open-air basking before 8:31 am, predictions in the time period part, the model coefficients are obtained by multiple linear
of 8:31 am to 8:49 am should be neglected because thermodynamic regression method from the aspect of statistics. The accuracy of the
characteristic of the SAC changed at 8:31am when the fan started. TFM is highly dependent on the sample data. The model fitting of
And the collector underwent a transitional state during that period. TFM does not consider the case of global irradiance below 300 W/
When twice the collector time constants (18 min) had elapsed since m2. Hence, the predictions at the moments 14:38, 14:50, 15:09, etc.
8:31 am, the collector thermal inertia gradually dominated the in Fig. 9 are poor compared to the experimental data, when the
momentary heat storage quantity of the collector. However, the instantaneous solar irradiance changes sharply below 300 W/m2.
collector's thermal inertia effect is neglected by the SST method. When it comes to the TICM, the variation trend of the prediction
Take the test requirements of SST method in the ASHRAE 93-2003 curve in Fig. 10 is basically in accordance with that of the experi-
standard [1] as an example. The steady-state tests shall be per- mental data. It suggests that, the TICM which considers collector
formed during periods when the sky is clear (i.e., substantially free thermal inertia correction can damp the effect of very sharp irra-
of clouds) such that the solar irradiance incident upon the aperture diation variations. Fig. 11 demonstrates the relative errors between
plane does not vary more than ±32 W/m2 for durations of 10 min or the predictions by the three models and the experimental data of
two time constants, whichever is greater, both prior to and during Qu . The maximum momentary relative error of Qu predicted by the
the period when data are taken. This strict test conditions remove SST model and the second-order differential TFM is larger than 40%,
the collector's thermal inertia effect from the test points of SST while that of Qu predicted by the TICM is around 20%. The averaged
model fitting. In the process of collector dynamic test under vari- overestimated percentage of the predictions by the SST model is
able weather conditions, the collector's thermal inertia effect re- 6.4%. And the averaged prediction errors by the TFM and the TICM
sults in a deviation from the steady-state. Hence, predictions by the are 2.7% and 1.3%, respectively. It suggests that the TICM performs
SST model generally overestimate the instantaneous useful heat better than the SST model and the second-order differential TFM in
gains in dynamic tests when there is good solar irradiance without
sudden decrease, as shown in most time of Fig. 8. While the solar
irradiance suddenly falls down, the stored heat in the collector due
to thermal capacitance transfers to the working fluid to damp a
sudden decrease of the collector outlet temperature. In this case,
predictions by the SST model underestimate the instantaneous
useful heat gains, as predictions at the moments 10:30, 11:20,
14:38, 14:50, 15:09, etc. in Fig. 8 when the instantaneous solar
irradiance sharply falls down. At the moments 14:38, 14:50, 15:09,
etc., the solar irradiance (see Fig. 6) falls down below the critical
radiation level, it is obvious that predictions by the SST model in

Table 1
Parameters obtained by steady-state test method.
·
Parameters FR ðtaÞen FR UL T mf Cf ðmCÞe Dt j
2   
Units e W/(m C) s W/ C J/ C s e
Fig. 9. Comparison of instantaneous Qu between the experimental data and the pre-
Values 0.495 5.271 540 41.205 44501.4 10 216
diction by TFM.
J. Deng et al. / Renewable Energy 76 (2015) 679e686 685

Besides, the coefficients in the TICM can be obtained based on


the SST method and a single test of the collector time constant T.
Thus the model not only pertains to flat-plate solar collectors, but
also pertains to other types of solar collectors if they can be tested
by SST method.

5. Conclusion

An analytical model in the form of series expansion is developed


based on the consideration of effective thermal capacitance of a
flat-plate solar collector. It works with a relatively small time in-
terval. The model reveals that instantaneous useful heat gain of a
solar collector at one moment consists of the steady-state useful
heat gain and corresponding thermal inertia correction. Instanta-
neous useful heat gain of a moment t is mainly affected by the
Fig. 10. Comparison of instantaneous Qu between the experimental data and the
collector thermal inertial effect during the time period of [t-3T, t],
prediction by TICM.
in which T is the response time constant of the collector.
It is validated by the experimental data that the TICM performs
the period of 8:49e16:24. The TICM can catch the momentary better than the SST model and the second-order differential TFM
thermal characteristics of instantaneous useful heat gain when the for predicting the momentary thermal characteristics of the SAC
instantaneous solar irradiance changes sharply at the moments when the solar irradiance changes sharply at one moment. The
14:38, 14:50, 15:09, etc. While the SST method and the TFM model also pertains to other types of solar collectors if they can be
perform poorly. Moreover, small errors also exist in the prediction tested by SST method.
by TICM. As mentioned in section 3.1, the system of a collector is an
n-order system and a first-order differential model is just a Acknowledgments
simplified model of the collector. And thus the thermal inertial
effect of the collector is not strictly an exponential relation. Hence, The financial support from the National Natural Science Foun-
the errors are mainly derived from the calculation deviation of the dation of China (Grant No.51038005) and the Doctoral Program of
collector thermal inertia effect. Higher Education Research Fund (Grant No.2011002110093) are
gratefully acknowledged. Besides, the authors wish to thank Pro-
fessor Zhifeng Wang of the Chinese Academy of Sciences for
4.3. Applicable conditions of the TICM
allowing us to use his test platform of solar air collectors.
The TICM given by Equation (18) is deduced by the hypothesis
that temperatures Tfi , Ta and absorbed solar irradiance S can be References
assumed as constants, so it works with a small time interval Dt,
[1] ANSI/ASHRAE Standard 93-2003. Methods of testing to determine the thermal
when the solar irradiances, ambient temperatures and collector
performance of solar collectors. Atlanta: ASHRAE Inc.; 2003.
inlet temperatures are measured or known. And the wind speed [2] Amer EH, Nayak JK, Sharma GK. Transient test methods for flat-plate collec-
·
should be small, e.g., within 0e3 m/s. The mass flow rate mf of the tors: review and experimental evaluation. Sol Energy 1997;60(5):229e43.
working fluid should be kept constant. When the mass flow rate [3] Nayak JK, Amer EH. Experimental and theoretical evaluation of dynamic test
procedures for solar flat-plate collectors. Sol Energy 2000;69(5):377e401.
changes, the factor FR in the SST method should be corrected in [4] Cooper PI, Dunkle RV. A non-linear flat-plate collector model. Sol Energy
·
terms of collector flow efficiency factor F 0 with changed mf , as 1981;26(2):133e40.
shown in Equation (20) [17]. [5] Chungpaibulpatana S, Exell RHB. The effects of using a one-node heat
capacitance model for determining solar collector performance parameters by
· " !# transient test methods. Sol Wind Technol 1988;5(4):411e21.
mf Cf A a UL F 0 [6] Chungpaibulpatana S, Exell RHB. A transient method to determine the per-
FR ¼ 1  exp  · (20) formance of solar collectors with very large fluid heat capacities. Int J Ambient
A a UL mf Cf Energy 1990;11(2):83e98.
[7] Perers B. Dynamic method for solar collector array testing and evaluation with
standard database and simulation programs. Sol Energy 1993;50(6):517e26.
[8] Perers B. An improved dynamic solar collector test method for determination
of non-linear optical and thermal characteristics with multiple regression. Sol
Energy 1997;59(4e6):163e78.
[9] EN 12975-2. Thermal solar systems and components e solar collectors e
part2: test methods. 2006.
[10] Fischer S, Heidemann W, Müller-Steinhagen H, Perers B, Bergquist P,
Hellstro€ m B. Collector test method under quasi-dynamic conditions according
to the European standard EN12975-2. Sol Energy 2004;76(1):117e23.
[11] Frid SE. Multinode models and dynamic testing methods of solar collectors.
Sol Wind Technol 1990;7(6):655e61.
[12] Wijeysundera NE, Hawlader MNA, Foong KY. Estimation of collector perfor-
mance parameters from daily system tests. J Sol Energy Eng Trans ASME
1996;118(1):30e6.
[13] Emery M, Rogers BA. On a solar collector thermal performance test method for
use in variable conditions. Sol Energy 1984;33(2):117e23.
[14] Prapas DE, Norton B, Milonidis E, Probert SD. Response function for solar-
energy collectors. Sol Energy 1988;40(4):371e83.
[15] Wang XA, Xu YF, Meng XY. A filter method for transient testing of collector
performance. Sol Energy 1987;38(2):125e34.
[16] Amer EH, Nayak JK, Sharma GK. A new dynamic method for testing solar flat-
Fig. 11. Comparison of the relative errors between the predictions by the three models plate collectors under variable weather. Energy Convers Manag 1999;40(8):
and the experimental data of Qu . 803e23.
686 J. Deng et al. / Renewable Energy 76 (2015) 679e686

[17] Duffie JA, Beckman WA. Solar engineering of thermal processes. 2nd ed. New [21] Kong WQ, Wang ZF, Li X, Li X, Xiao N. Theoretical analysis and experimental
York: John Wiley and Sons; 1991 [Chapter 6]. verification of a new dynamic test method for solar collectors. Sol Energy
[18] Kong WQ, Wang ZF, Fan JH, Bacher P, Perers B, Chen ZQ, et al. An improved 2012;86(1):398e406.
dynamic test method for solar collectors. Sol Energy 2012;86(6):1838e48. [22] Hou HJ, Wang ZF. A new method for the measurement of solar collector time
[19] Hou HJ. A transient test method for thermal performance of flat-plate solar constant. Renew Energy 2005;30(6):855e65.
collectors. Acta Energiae Solaris Sin 2004;25(3):310e4. [23] ISO. Guide to the expression of uncertainty in measurement. Switzerland:
[20] Xu L, Wang ZF, Yuan GF, Li X, Ruan Y. A new dynamic test method for thermal International Organization for Standardization; 1995.
performance of all-glass evacuated solar air collectors. Sol Energy 2012;86(5):
1222e31.

Das könnte Ihnen auch gefallen